diff --git a/app/src/git_commit_generator/ollama_client.py b/app/src/git_commit_generator/ollama_client.py
new file mode 100644
index 0000000..4db8ad8
--- /dev/null
+++ b/app/src/git_commit_generator/ollama_client.py
@@ -0,0 +1,144 @@
+ """Ollama client wrapper for LLM interactions."""
+import ollama as ollama_lib
+from ollama import ChatResponse, ListResponse
+
+
+class OllamaClient:
+    """Client for interacting with Ollama LLM."""
+
+    def __init__(self, host: str = "http://localhost:11434", model: str = "llama3"):
+        """Initialize Ollama client.
+
+        Args:
+            host: Ollama server URL.
+            model: Default model to use.
+        """
+        self.host = host
+        self.model = model
+        ollama_lib.host = host
+
+    def check_connection(self) -> bool:
+        """Check if Ollama is running and accessible.
+
+        Returns:
+            True if connection successful, False otherwise.
+        """
+        try:
+            response: ListResponse = ollama_lib.list()
+            return True
+        except Exception:
+            return False
+
+    def check_model_available(self, model: str) -> bool:
+        """Check if a specific model is available.
+
+        Args:
+            model: Model name to check.
+
+        Returns:
+            True if model is available, False otherwise.
+        """
+        try:
+            response: ListResponse = ollama_lib.list()
+            return any(m["name"] == model or m["name"].startswith(model)
+                      for m in response.get("models", []))
+        except Exception:
+            return False
+
+    def generate_commit_message(
+        self, diff: str, prompt: str, model: str | None = None
+    ) -> str:
+        """Generate a commit message from diff.
+
+        Args:
+            diff: Git diff to analyze.
+            prompt: Prompt template to use.
+            model: Optional model override.
+
+        Returns:
+            Generated commit message.
+        """
+        response: ChatResponse = ollama_lib.chat(
+            model=model or self.model,
+            messages=[
+                {
+                    "role": "system",
+                    "content": "You are a helpful assistant that generates git commit messages following conventional commit format. Be concise and descriptive.",
+                },
+                {
+                    "role": "user",
+                    "content": f"{prompt}\n\nGit diff to analyze:\n\n{diff}",
+                },
+            ],
+            stream=False,
+        )
+        return response["message"]["content"]
+
+    def generate_changelog(
+        self, commits: str, prompt: str, model: str | None = None
+    ) -> str:
+        """Generate changelog from commits.
+
+        Args:
+            commits: Commit history to analyze.
+            prompt: Prompt template to use.
+            model: Optional model override.
+
+        Returns:
+            Generated changelog.
+        """
+        response: ChatResponse = ollama_lib.chat(
+            model=model or self.model,
+            messages=[
+                {
+                    "role": "system",
+                    "content": "You are a helpful assistant that generates changelogs from git commit history. Group changes by type and create a well-structured markdown changelog.",
+                },
+                {
+                    "role": "user",
+                    "content": f"{prompt}\n\nCommit history to analyze:\n\n{commits}",
+                },
+            ],
+            stream=False,
+        )
+        return response["message"]["content"]
+
+    def list_models(self) -> list[dict]:
+        """List available Ollama models.
+
+        Returns:
+            List of available models with their details.
+        """
+        try:
+            response: ListResponse = ollama_lib.list()
+            return response.get("models", [])
+        except Exception:
+            return []
+
+    def pull_model(self, model: str) -> bool:
+        """Pull a model from Ollama registry.
+
+        Args:
+            model: Model name to pull.
+
+        Returns:
+            True if successful, False otherwise.
+        """
+        try:
+            ollama_lib.pull(model)
+            return True
+        except Exception:
+            return False
+
+
+def get_ollama_client(host: str = "http://localhost:11434", model: str = "llama3") -> OllamaClient:
+    """Get OllamaClient instance.
+
+    Args:
+        host: Ollama server URL.
+        model: Default model to use.
+
+    Returns:
+        OllamaClient instance.
+    """
+    return OllamaClient(host=host, model=model)